Error Handling - Cryptography in C and C++ - page 356

Table 16-1. LINT function error codes

Code Value Explanation

E_LINT_OK

0x0000 everything ok

E_LINT_EOF

0x0010 file I/O error in the stream operator <<

or >>

E_LINT_DBZ 0x0020 division by zero

E_LINT_NHP Heap error: new returns the NULL pointer 0x0040

E_LINT_OFL

0x0080 overflow in function or operator

E_LINT_UFL 0x0100 underflow in function or operator

E_LINT_VAL 0x0200 an argument of a function is uninitialized

or has an illegal value

E_LINT_BOR 0x0400 incorrect base passed as argument to a

constructor

E_LINT_MOD even modulus in mexpkm()

0x0800

E_LINT_NPT 0x1000 NULL pointer passed as argument

E_LINT_RIN 0x2000 call to an uninitialized pseudorandom number generator

16.2 User-Defined Error Handling

As a rule, it is necessary to adapt error handling to particular requirements. The
LINT class offers support in this regard in that the LINT error function panic()
can be replaced by user-defined functions. Additionally, the following function is
called, which takes as argument a pointer to a function:

void
LINT::Set_LINT_Error_Handler (void (*Error_Handler)
(LINT_ERRORS, const char*, int, int, const, char*))

{

LINT_User_Error_Handler = Error_Handler;

}

The variable LINT_User_Error_Handler is defined and initialized in
flintpp.cpp as

static void (*LINT_User_Error_Handler)
(LINT_ERRORS, const char*, int, int, const char*) = NULL;

If this pointer has a value other than NULL , then the specified function is called
instead of panic() , and it contains the same information as panic() would have.
With respect to the implementation of a user-defined error-handling routine
one has a great deal of freedom. But one must realize that the errors reported by
the class LINT usually signal program errors, which are irreparable at run time. It